Designing Productive Assembly System Configurations Based on Hierarchical Subassembly Decomposition with Application to Automotive Battery Packs.

Abstract

This thesis develops a systematic method to design assembly systems with hybrid configurations by considering the assembly hierarchy associated with product designs and their varieties and applies it for automotive battery packs. With the growing concern of fossil fuel depletion and climate change, lithium-ion batteries are being widely adopted in personal transportation systems. A battery pack usually has a hierarchical composition of components assembled in some repetitive patterns. A lot of battery designs are emerging on the market. They require different processes and equipment from cell to module assembly, but similar processes and equipment from module to pack assembly. Conventional assembly system with a serial configuration has limitations in coping with increasing demand and fast development of the battery products. There is a strong need to develop assembly systems with complex, non-serial (hybrid) configurations to deal with the challenges, e.g. a system layout with multiple branch lines that converge to a common assembly line. Such configurations could be asymmetric and allows for pre-assembly of different components on multiple lines simultaneously, thereby potentially enhancing the system throughput and reconfigurability, while effectively dealing with product variety. Previous research has been focused on sequential task sequence generation but did not address the impact of product assembly hierarchy on configuration. Limited work exists addressing the line balancing problem on complex configuration. There is also a lack of research on non-serial system configuration design for both known and future product variants. Existing methods for designing complex system configuration do not consider equipment selection. Based on graph theory and combinatorial mathematics, a new algorithm for analyzing the liaison topographic patterns in products is developed to identify optimal assembly/subassembly decompositions that link product designs to system configurations. Compared with the sequential method for system design, the integrated approach of concurrent assembly process planning, system balancing, equipment selection, and system configuration design leads to higher throughput performances. Meanwhile, a method is developed to model the impact of product variety on system configuration design by considering stochastic product mix changes. This research enhances the understanding of the complex interactions among product designs, product varieties, and assembly system configurations.